Hi The OCXO vs OCXO stuff looks about as good as you will get until they both warm up for a few weeks. The output of the TBolt has some sort of problem that can be troubleshot. The key data there is the phase plot rather than either of the other two. You know what a “good” phase plot looks like from the OCXO vs OCXO stuff.
The objective in your troubleshooting is to figure out what’s wrong with the TBolt. Monitoring logs with Lady Heather (that software is one of the best things about the TBolt) would be the first thing to do. See when you get the spikes and see what’s going on with the unit. Another approach would be to toss a second or third GPSDO into the mix or an Rb. They would not tell you a lot in the troubleshooting. They would give you some different curves to look at. Since the data only starts to appear at 1,000 seconds, things take a while. You *should* have a hundred or so samples at the “start” (in this case 1K sec). That gets you out to a minimum run of a day and a “two decades” run of many days. Back in the old days we spent the $50 to build a single mixer setup in these cases. None of us had the patience for one day runs …. The most interesting way to do it is one that Bruce suggested several years back. I don’t think anybody has tried it and reported back (yet). Rather than take the double balanced mixer output and limit it, you lowpass filter it and feed it into an ADC. DSP-ish software extracts the frequency data from the estimated zero crossings of the beat note. Doing it with a normal sound card isn’t very easy since the have issues (noise etc) below 20 Hz or so. Bob > On May 30, 2015, at 11:11 PM, Vlad <t...@patoka.org> wrote: > > > I did couple of other measurements, using MV89A as a ref source for 5386A > > DATUM OCXO is DUT: > http://www.patoka.ca/OCXO/OCXO-MV89Aref-DATUM-AD.png > http://www.patoka.ca/OCXO/OCXO-MV89Aref-DATUM-TM.png > http://www.patoka.ca/OCXO/OCXO-MV89Aref-DATUM-FD.png > Drift(Hz/sec): -2.20E-7 > > Trimble TB is DUT: > http://www.patoka.ca/OCXO/OCXO-MV89Aref-TB-AD.png > http://www.patoka.ca/OCXO/OCXO-MV89Aref-TB-TM.png > http://www.patoka.ca/OCXO/OCXO-MV89Aref-TB-FD.png > Drift(Hz/sec): -5.17E-8 > > Looking to the drift values, the case with T-Bolt has better value, I think. > But it has bigger spikes. > > > > On 2015-05-29 18:59, Bob Camp wrote: >> Hi >> I think that looks about as good as your counter can do. The slope down at >> the start is the resolution of the counter. The slope up on the right >> is the actual stability of >> the oscillator. >> Bob >>> On May 29, 2015, at 9:52 AM, Vlad <t...@patoka.org> wrote: >>> Hello, >>> I've finished the test. Here is the links to charts: >>> http://www.patoka.ca/OCXO/OCXO-8663ref-MV89A-AD.png >>> http://www.patoka.ca/OCXO/OCXO-8663ref-MV89A-PD.png >>> http://www.patoka.ca/OCXO/OCXO-8663ref-MV89A.png >>> It looks different, but its not better, I think. >>> Regards, >>> Vlad >>> On 2015-05-28 23:21, Vlad wrote: >>>> Bob, >>>> You are correct. I am using HP 5386A. And its only HP counter I have. >>>> I set Gate time equal to 10sec. May be this gate time is an issue ? >>>> The other settings is .1s, and 1s. >>>> Meantime, I did switch ref signal source from Trimble TB to >>>> Oscilloquarz 8663-SX OCXO and connected MV89A to counter input. 8 >>>> hours later, I'll see what will be on chart. >>>> Regards, >>>> Vlad >>>> On 2015-05-28 21:50, Bob Camp wrote: >>>>> Hi >>>>> *IF* I’m understanding you correctly, you are measuring this data >>>>> with an HP 5386. >>>>> That’s only a 4 ns counter. I agree with Tom that there is something >>>>> weird in your setup, but that >>>>> still does not help the basic limit of the counter’s resolution. You >>>>> will need to be out around 1,000 >>>>> seconds or more to get to the noise on the OCXO. At any time shorter >>>>> than that you are either >>>>> measuring the counter’s noise or noise in your setup. >>>>> If you have one, a 5335 would be a better counter to use than the 5386. >>>>> Bob >>>>>> On May 27, 2015, at 10:04 PM, Vlad <t...@patoka.org> wrote: >>>>>> Hello, >>>>>> Recently, I acquired video distribution amplifier (model Altinex >>>>>> DA1804NT). It says it could handle 350Mhz. I am planing to use it to >>>>>> split 10Mhz Ref signal from my OCXO. This OCXO I bought on auction site. >>>>>> It has Oscilloquarz model 8663-SX plus some components which give me 4 >>>>>> outputs - 2 sines and two squares. After few days of "warm up" I did >>>>>> measurement of ADV from one of the sine outputs and then measurement of >>>>>> the same sine signal from OCXO, but this time through the Altinex DA. >>>>>> Each test took 8 hours. Attached is the results. >>>>>> NOTE: on the ALTINEX* charts I forgot to change the label for OCXO. So, >>>>>> its not DATUM. It is Oscilloquarz 8663-SX >>>>>> I was using Trimble TB as a ref signal for HP 5386 >>>>>> Since I have no instruments which measuring phase noise, I am >>>>>> interesting, what will be the quick method to measure the phase noise >>>>>> for that DA ? >>>>>> -- >>>>>> WBW, >>>>>> V.P.<ALTINEX-DA.png><ALTINEX-DA-F.png><ALTINEX-DA-T.png><OCXO2.png>_______________________________________________ >>>>>> time-nuts mailing list -- time-nuts@febo.com >>>>>> To unsubscribe, go to >>>>>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >>>>>> and follow the instructions there. >>> -- >>> WBW, >>> V.P. >>> _______________________________________________ >>> time-nuts mailing list -- time-nuts@febo.com >>> To unsubscribe, go to >>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >>> and follow the instructions there. >> _______________________________________________ >> time-nuts mailing list -- time-nuts@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >> and follow the instructions there. > > -- > WBW, > > V.P. > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. _______________________________________________ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
